Re: Introduction of libacpica to get pci irqs and shutdown

Hi On 4/9/22 1:47 pm, Damien Zammit wrote:> This patch series has been tested to boot a rumpdisk > and fetch the PCI IRQs of disk controllers from acpi.static translator > via bootstrapping acpi -> pci -> rumpdisk -> ext2fs -> exec. On native hardware, acpi translator fails with intnull(9) and h

[PATCH 5/5 rumpkernel] pci-userspace: Add acpi lookup of irqs

This depends on having installed. It will now fail to run if acpi translator is not running as it needs the actual pci irq. --- debian/patches/acpi.diff | 74 debian/patches/series| 1 + 2 files changed, 75 insertions(+) create mode 100644 debian/p

[PATCH 3/5] acpi: Convert translator to an emulated mach device

--- acpi/Makefile | 12 +-- acpi/acpifs.c | 34 +--- acpi/acpifs.h | 6 +- acpi/main.c| 206 +++-- acpi/options.c | 20 + acpi/options.h | 8 ++ 6 files changed, 243 insertions(+), 43 deletions(-) diff --git a/acpi/Makefile b/acpi/

[PATCH 4/5] acpi: Remove references to /dev/mem

--- acpi/acpi.c | 109 1 file changed, 34 insertions(+), 75 deletions(-) diff --git a/acpi/acpi.c b/acpi/acpi.c index 9827232a..dd32d60f 100644 --- a/acpi/acpi.c +++ b/acpi/acpi.c @@ -30,37 +30,13 @@ #include "myacpi.h" -int -mmap_phys_acpi_

[PATCH 1/5] acpi: Link translator to libacpica and provide RPCs

Provides two new acpi RPCs to sleep the machine and to get the irq of any pci device. ACPI mode is enabled by default when the translator is started. NB: Merging this commit means libacpica is a build dep. --- acpi/Makefile | 19 +++-- acpi/acpi-ops.c | 83 +

[PATCH 2/5] shutdown: Use new acpi RPC to halt machine, clean up

This allows clean shutdown of all modern x86 machines (not just qemu) by using the acpi translator to call into libacpica code. --- shutdown/Makefile| 2 +- shutdown/acpi_shutdown.c | 66 shutdown/acpi_shutdown.h | 18 --- shutdown/shutdown

Introduction of libacpica to get pci irqs and shutdown

Hi! This patch series has been tested to boot a rumpdisk and fetch the PCI IRQs of disk controllers from acpi.static translator via bootstrapping acpi -> pci -> rumpdisk -> ext2fs -> exec. It also cleanly shuts down the machine by sending an RPC to acpi to call the new shutdown method inside li

Re: Freeze booting from upstream gnumach

NOTE: The libc6 version in Debian testing is 2.34-7 El sáb, 3 sept 2022 a las 19:32, Almudena Garcia (< liberamenso10...@gmail.com>) escribió: > Hi all: > > By some reason, after upgrade the host machine (Debian testing) the > problems seems solved. > Now gnumach boots without freeze > > El sáb,

Re: Freeze booting from upstream gnumach

Hi all: By some reason, after upgrade the host machine (Debian testing) the problems seems solved. Now gnumach boots without freeze El sáb, 3 sept 2022 a las 18:33, Almudena Garcia (< liberamenso10...@gmail.com>) escribió: > apt policy shows libc0.3's version is 2.34-8~0 > > El sáb, 3 sept 2022

Re: Freeze booting from upstream gnumach

apt policy shows libc0.3's version is 2.34-8~0 El sáb, 3 sept 2022 a las 1:53, Samuel Thibault () escribió: > Almudena Garcia, le ven. 02 sept. 2022 21:24:44 +, a ecrit: > > Yes, I full-upgraded to latest > > "latest" remains relative. > > Actual versions of hurd, libc0.3 will help to make su

Re: sigdescr_np for SIGEMT is useless

Samuel Thibault wrote: > the foo.c testcase is missing as attachment Oops, sorry. > but please check whether this is a hurd-specific > issue or a generic glibc issue, as we have really much less manpower > than the glibc team in general. Here it is a bug in glibc in general, > and in Linux too (e

Re: thrd_current crashes in the main thread

Samuel Thibault wrote: > > The ISO C function thrd_current() crashes in the main thread on glibc/Hurd, > That's already fixed in 2.35. Fix now also pending for 2.34-8. Thanks! Cool.

Re: strerrorname_np returns NULL for some defined errno values

Hello, Bruno Haible, le sam. 03 sept. 2022 08:28:31 +0200, a ecrit: > The function strerrorname_np is supposed to return non-NULL for all defined > errno values. [1] > > On Hurd, for the errnos with names > ED_* > EKERN_* > EMACH_* > EMIG_* > it returns NULL. > > How to reproduce: > $ gc

Re: Freeze booting from upstream gnumach

Damien Zammit, le sam. 03 sept. 2022 04:23:48 +, a ecrit: > On 3/9/22 1:26 am, Almudena Garcia wrote: > > I've just compiled gnumach from upstream sources, and I noticed that there > > are a freeze in the booting, just in the pci-arbitrer starting. > > > > I attach a screenshot with the proble

Re: thrd_current crashes in the main thread

Hello, Bruno Haible, le sam. 03 sept. 2022 08:32:32 +0200, a ecrit: > The ISO C function thrd_current() crashes in the main thread on glibc/Hurd, > but not on glibc/Linux. > > How to reproduce: > $ gcc -Wall foo.c > $ ./a.out > crashes with SIGSEGV. > > Seen on Debian GNU/Hurd, with glibc 2.34.

Re: sigdescr_np for SIGEMT is useless

Hello, It is nice to see precise bug reports (though here the foo.c testcase is missing as attachment), but please check whether this is a hurd-specific issue or a generic glibc issue, as we have really much less manpower than the glibc team in general. Here it is a bug in glibc in general, and in